A sample is biased if its collection method systematically makes some members of the population more or less likely to be represented. Identify bias by comparing the target population with the sampling frame, then checking selection, non-response, and data-collection methods.
How to Check for Bias
- Define the target population.
- Identify the sampling frame: the group from which the sample was chosen.
- Ask whether each relevant group had a reasonable chance of selection.
- Check whether non-response, voluntary participation, or collection methods could systematically affect results.
- State which groups are overrepresented or underrepresented and how this may distort the conclusion.
| Source of bias | Identification | Example |
|---|---|---|
| Convenience sampling | Participants are chosen because they are easy to reach | Surveying only library users about student study time |
| Undercoverage | Part of the population is absent from the sampling frame | An online survey excludes people without internet access |
| Voluntary-response bias | People choose to participate, so strong opinions may dominate | An optional poll about school lunches |
| Non-response bias | Selected people do not reply, and non-responders may differ | Students with low attendance ignore an attendance survey |
| Poor question design | Leading or ambiguous wording influences answers | Asking, “Do you agree that the excellent timetable should remain?” |
A simple random sample or appropriately designed stratified sample usually reduces sampling bias because selection is based on chance and relevant groups can be represented proportionally. However, random selection does not remove non-response or measurement bias.